A virtual machine for sensor networks

Abstract

Sensor networks are increasingly being deployed for a wide variety of tasks. Today, in these networks, the development, deployment, and maintenance of applications are performed largely ad-hoc. Existing platforms help somewhat but also introduce implicit trade-offs. In one extreme, low-level programming platforms and languages make programming cumbersome and error-prone. In the other extreme, declarative approaches greatly facilitate programming but restrict what can be done. In both cases, additional limitations include lack of support for concurrency, difficulties in changing applications, and insufficient abstractions from low-level details. This paper presents SwissQM, a virtual machine designed to address all these limitations. SwissQM offers a platform-independent programming abstraction that is geared towards data acquisition and in-network data processing.